2nd Quarter 2008

U.S. Housing

Market Conditions
August 2008

During the second quarter of 2008, builders took out

permits for new housing at 1,033,000 units (SAAR),

up 4 percent from the first quarter but down 30 percent
from the second quarter of 2007. Single-family permits
were issued for 633,000 (SAAR) housing units, a
decrease of 2 percent from the first quarter and a
decrease of 40 percent from the second quarter of
2007. This decrease is the 11th consecutive
quarterly decline for single-family permits.

Builders started construction on 1,016,000 (SAAR)

new housing units in the second quarter of 2008,

down 4 percent from the first quarter and down 30

Manufactured housing shipments continued at very

low levels. In the second quarter of 2008, manufacturers shipped 88,000 (SAAR) housing units, down
4 percent from the first quarter and down 11 percent
from the second quarter of 2007.

Housing Marketing
Housing sales and builders attitudes continued downward in the second quarter of 2008, but sales prices
showed signs of strength. Although new home sales
have declined in the past 11 quarters, and existing
home sales have fallen for five consecutive quarters,
the median price of new homes was unchanged and

Contents.................................. 2
Using HMDA and Income
Leverage To Examine Current
Mortgage Market Turmoil........ 4
National Data......................... 14
Regional Activity................ 28
Historical Data.. ...................64
G
SIN
OU

AN

U.S. Department of Housing and Urban Development

Office of Policy Development and Research

MENT OF
RT
H
PA

Most housing production indicators declined in the

second quarter of 2008 but reflected a slight improvement
over the universal declines recorded during the past
several quarters. The number of building permits issued
increased, although starts and completions continued
to decline. Manufactured housing has posted nearly
continuous declines since the hurricane-induced orders
of late 2005. Shipments of manufactured homes are now
below 90,000 units at a seasonally adjusted annual rate
(SAAR), the lowest since the third quarter of 1961.

units in the second quarter of 2008, down 11 percent

from the first quarter and down 27 percent from the
second quarter of 2007. This decrease is the ninth
consecutive quarterly decline. Single-family completions totaled 851,000 (SAAR) in the second quarter
of 2008, down 9 percent from the first quarter and
down 33 percent from the second quarter of 2007,
reflecting the ninth consecutive quarterly decline
for this indicator.

RB

EN

Housing Production

Builders completed 1,118,000 (SAAR) new housing

The housing market had a poor second quarter in 2008,

continuing 2 years of decline. The number of singlefamily building permits, starts, and completions
declined in the second quarter, as did new and existing
home sales. Excessive inventories of both new and
existing homes remain, enough to last 10 to 11 months.
The multifamily sector is somewhat mixed: permits
and starts increased, but completions decreased. The
subprime meltdown continues, with foreclosure rates
on subprime adjustable-rate mortgages posting a
20-percent increase over the previous quarter. Conditions
in the rental housing market showed little change
from the first quarter of 2008, with a tiny decrease in
the vacancy rate and no change in absorptions. The
overall economy posted a Gross Domestic Product
(GDP) growth rate of 1.9 percent in the second quarter
of 2008. The housing component of GDP decreased by
15.6 percent, leading to a reduction of GDP growth by
0.62 percentage point.

percent from the second quarter of 2007. Singlefamily housing starts totaled 670,000 (SAAR) housing
units, down 8 percent from the first quarter and
down 42 percent from the second quarter of 2007.
This drop is the ninth consecutive quarterly decline
for single-family starts.

U.S.
DE

Summary

AN

D EVEL

OP

Contents
Summary.......................................................1
Housing Production................................. 1
Housing Marketing.................................. 1
Affordability and Interest Rates............... 3
Multifamily Housing............................... 3
Using HMDA and Income Leverage To
Examine Current Mortgage Market
Turmoil................................................... 4
Notes......................................................... 13
Reference.................................................. 13

National Data...................................... 14
Housing Production....................................14
Permits...................................................... 14
Starts......................................................... 15
Under Construction.................................15
Completions............................................16
Manufactured (Mobile) Home
Shipments.............................................16
Housing Marketing..................................... 17
Home Sales............................................... 17
Home Prices.............................................18
Housing Affordability.............................. 19
Apartment Absorptions...........................20
Manufactured (Mobile) Home
Placements............................................. 20
Builders Views of
Housing Market Activity.....................21
Housing Finance.........................................22
Mortgage Interest Rates...........................22
FHA 14 Family Mortgage
Insurance...............................................23
PMI and VA Activity...............................23
Delinquencies and Foreclosures..............24
Housing Investment...................................25
Residential Fixed Investment and
Gross Domestic Product.......................25
Housing Inventory......................................26
Housing Stock.......................................... 26
Vacancy Rates........................................... 27
Homeownership Rates............................. 27

Regional Activity.............................. 28
Regional Reports........................................28
New England............................................ 28
New York/New Jersey............................. 30
Mid-Atlantic............................................. 31
Southeast/Caribbean................................33
Midwest...................................................35
Southwest................................................. 37
Great Plains.............................................38
Rocky Mountain...................................... 40
Pacific.......................................................41
Northwest................................................. 43
Housing Market Profiles............................45
Albuquerque, New Mexico..................... 45
Austin-Round Rock, Texas...................... 46
Bremerton-Silverdale, Washington......... 47

Summary

Chattanooga, Tennessee-Georgia............ 48
Dallas-Plano-Irving, Texas....................... 49
Lawrence, Kansas..................................... 51
Mobile, Alabama...................................... 52
Philadelphia, Pennsylvania...................... 53
Phoenix, Arizona...................................... 55
Sacramento--Arden-Arcade--Roseville,
California............................................... 56
San Antonio, Texas.................................. 57
Spartanburg, South Carolina...................59
St. George, Utah......................................60
Units Authorized by Building Permits,
Year to Date: HUD Regions
and States............................................... 62
Units Authorized by Building Permits,
Year to Date: 50 Most Active Core
Based Statistical Areas (Listed by
Total Building Permits)........................63

Historical Data...................................64
Table 1
New Privately Owned Housing
Units Authorized: 1967Present.......... 64
Table 2
New Privately Owned Housing
Units Started: 1967Present.................65
Table 3
New Privately Owned Housing Units
Under Construction: 1970Present......66
Table 4
New Privately Owned Housing
Units Completed: 1970Present...........67
Table 5
Manufactured (Mobile) Home
Shipments, Residential Placements,
Average Prices, and Units for Sale:
1977Present.........................................68
Table 6
New Single-Family Home Sales:
1970Present.........................................69
Table 7
Existing Home Sales: 1969Present.....70
Table 8
New Single-Family Home Prices:
1964Present.......................................... 71
Table 9
Existing Home Prices: 1969Present....72
Table 10
Repeat Sales House Price Index:
1991Present............................................73
Table 11
Housing Affordability Index:
1973Present............................................74
Table 12
Market Absorption of New Rental
Units and Median Asking Rent:
1970Present.........................................75
Table 13
Builders Views of Housing Market
Activity: 1979Present.........................76

Table 14
Mortgage Interest Rates, Average
Commitment Rates, and Points:
1973Present.........................................77
Table 15
Mortgage Interest Rates, Fees,
Effective Rates, and Average Term
to Maturity on Conventional Loans
Closed: 1982Present.......................... 78
Table 16
FHA, VA, and PMI 14 Family
Mortgage Insurance Activity:
1971Present.........................................79
Table 17
FHA Unassisted Multifamily
Mortgage Insurance Activity:
1980Present.........................................80
Table 18
Mortgage Delinquencies and
Foreclosures Started: 1986Present......81
Table 19
Expenditures for Existing Residential
Properties: 19772007...........................82
Table 20
Value of New Construction Put in
Place, Private Residential Buildings:
1974Present.........................................83
Table 21
Gross Domestic Product and
Residential Fixed Investment:
1960Present.........................................84
Table 22
Net Change in Number of House holds by Age of Householder:
1971Present.......................................... 85
Table 23
Net Change in Number of House holds by Type of Household:
1971Present.........................................86
Table 24
Net Change in Number of House holds by Race and Ethnicity of
Householder: 1971Present................. 87
Table 25
Total U.S. Housing Stock:
1970Present....................................... 88
Table 26
Rental Vacancy Rates:
1979Present............................................89
Table 27
Homeownership Rates by Age
of Householder: 1982Present............. 90
Table 28
Homeownership Rates by Region and
Metropolitan Status: 1983Present......91
Table 29
Homeownership Rates by Race and
Ethnicity: 1983Present........................92
Table 30
Homeownership Rates by Household
Type: 1983Present...............................93

the price of existing homes rose in the second quarter.

Average prices for both new and existing homes rose.
At the end of the second quarter, inventories of homes
available for sale were sufficient to last for the next 10
to 11 months at the current sales rates. The nearly
continuous drop in new home sales is the likely source
of pessimism among builders as measured by the National
Association of Home Builders/Wells Fargo Housing
Market Index, which fell again in the second quarter.

Home builders were slightly more pessimistic in the

second quarter of 2008. The National Association of

Home Builders/Wells Fargo composite Housing
Market Index was 19 in the second quarter of 2008,
down 1 index point from the first quarter and down
11 index points from the second quarter of 2007.
The index is based on three componentscurrent
sales expectations, future sales expectations, and
prospective buyer traffic. Of these components,
future sales expectations and prospective buyer
traffic rose, but these increases were not enough to
offset the decrease in current sales expectations.

In the second quarter of 2008, 535,000 (SAAR) new

single-family homes were sold, down 5 percent from

the 561,000 (SAAR) homes sold in the first quarter
and down 37 percent from the second quarter of 2007.

Affordability and Interest Rates

REALTORS sold 4,913,000 (SAAR) existing single-

family homes in the second quarter of 2008, down

1 percent from the first quarter and down 16 percent
from the second quarter of 2007.

Housing affordability declined in the second quarter of

2008, according to the index published by the NATIONAL
ASSOCIATION OF REALTORS. The composite index
indicates that the family earning the median income
had 125.2 percent of the income needed to purchase the
median-priced, existing single-family home using standard
lending guidelines. This value is down 7.2 points from
the first quarter but up 16.0 points from the second
quarter of 2007. The decrease in affordability from the
first quarter is attributed to the 5-percent increase in
the median price of an existing single-family home and
the 11-basis-point increase in the mortgage interest
rate and was partially offset by an 2.2-percent increase
in median family income. The second quarter homeownership rate was 68.1, up 0.3 percentage point from
the first quarter rate but 0.1 percentage point below
the rate of the second quarter of 2007.

The median price of new homes sold in the second

quarter of 2008 was $234,100, unchanged from the

first quarter but down 3 percent from the second
quarter of 2007. The average price for new homes
sold in the second quarter was $304,700, up 5 percent
from the first quarter but down 2 percent from the
second quarter of 2007. A constant-quality house
would have sold for $303,500 in the second quarter,
up 3 percent from the first quarter but down 3 percent from the second quarter of 2007.

The NATIONAL ASSOCIATION OF REALTORS

reported that the median price for existing homes

was $208,100 in the second quarter of 2008, up 5
percent from the first quarter but down 7 percent
from the second quarter of 2007. The average price
for existing homes in the second quarter was
$252,400, up 3 percent from the first quarter but
down 7 percent from the second quarter of 2007.

Multifamily Housing
The multifamily (five or more units) housing sector
performed better than the single-family sector in the
second quarter of 2008. Production indicators were
mixed; building permits and starts increased, but
completions decreased. The absorption of new rental
units was unchanged, but the rental vacancy rate fell.

At the end of the second quarter of 2008, 426,000

new homes were in the unsold inventory, down 9

percent from the end of the first quarter and down
22 percent from the end of the second quarter of
2007. This inventory will support 10.0 months of
sales at the current sales pace, down 1.2 months
from the end of the first quarter but up 1.7 months
from the end of the second quarter of 2007. The
inventory of existing homes available for sale at the
end of the second quarter consisted of 4,490,000
homes, up 9 percent from the end of the first quarter
and up 3 percent from the end of the second quarter
of 2007. This inventory would last for 11.1 months
at the current sales rate, up 1.1 months from the
end of the first quarter and up 2.0 months from the
end of the second quarter of 2007.

In the second quarter of 2008, builders took out

permits for 364,000 new multifamily units, up

21 percent from the first quarter but down 1 percent
from the second quarter of 2007.

Construction was started on 329,000 new multifamily

units in the second quarter of 2008, up 9 percent

from the first quarter and up 24 percent from the
second quarter of 2007.

Builders completed 243,000 units in the second

quarter of 2008, down 17 percent from the first

quarter but up 10 percent from the second quarter
of 2007.

Summary

 The rental vacancy rate in the second quarter of 2008

completed in the first quarter of 2008 leased in the

first 3 months following completion. This absorption
rate is unchanged from the previous quarter but up
8 points from the rate recorded in the first quarter
of 2007.

was 10.0 percent, down 0.1 percentage point from

the first quarter but up 0.5 percentage point from
the second quarter of 2007.

Market absorption of new rental apartments was

unchanged, with 59 percent of new apartments

Using HMDA and Income Leverage To

Examine Current Mortgage Market Turmoil
Much attention is currently being devoted to understanding the nature and dimensions of the current
mortgage market turmoil. Most analyses rely on proprietary data rich in detail about specific loan terms
but with little demographic information on borrowers
and mortgage holders. In this article, we use Home
Mortgage Disclosure Act (HMDA) data and introduce
a measure of leverage and payment risk to examine the
dimensions of the current turmoil.1

magnitude of the current-day upheaval in our mortgage

and real estate markets. This article begins that effort
with a preliminary examination, followed by some
basic findings.
With the inclusion of both the loan amount (M) and the
borrowers gross income (Y) on which the lender relied
for qualifying the borrower, the mortgage-to-income
(M/Y) ratio can be formed. The M/Y ratio is a direct
measure of the income leverage employed by the borrower to obtain the mortgage loan. The higher the
ratio is, the more dollars of loan provided by the lender
per dollar of qualifying income. Moreover, a borrower
has only three ways to increase the leverage of his or
her income with virtually all mortgage products; that
is, a borrower can qualify for a higher loan amount
with a given income by (1) increasing the front-end
payment-to-income ratio that governs the size of the
payment and associated mortgage allowed, (2) lowering
the interest rate used to calculate the initial qualifying
payment and associated mortgage amount, or (3) reducing the rate at which principal is repaid by extending
the term or paying interest only, assuming property tax
rates and homeowner insurance premiums are fixed.

Data about mortgage applications and applicants together

with subsequent outcomes have been collected since
1990 and made available by the Federal Reserve Board
of Governors in cooperation with the Federal Financial
Institutions Examination Council. The data have been
used for myriad applications, including tracking trends
in the mortgage market and screening for compliance
with legislation and regulations intended to promote
fair lending and equal opportunity for securing credit.
Although much has been accomplished with HMDA
data, a perennial lament has been the lack of basic
underwriting variables that would permit analytic controls for collateral, credit, and payment risk. Standard
underwriting data, such as the loan-to-value (LTV) ratio,
property price, or appraisal, along with the borrowers
credit score and payment-to-income or debt-to-income
ratios, are not collected. The addition of these data
would allow for much more direct comparison of like
borrowers and loan terms, thereby enabling analysts to
better monitor trends in the character and quality of
mortgage lending and more directly identify situations
of noncompliance with fair housing and equal credit
opportunity regulations.

These findings follow directly from relating the standard self-amortizing mortgage formula for the monthly
principal, interest, tax, and insurance (PITI) payment
to the standard front-end ratio (FER) underwriting limitation to that payment. Thus,
(FER * Y)/12 = PITI = M*[An + (1/LTV)*(T + I)/12]

Where FER = front-end ratio,

Y = gross annual income,

PITI = monthly payment for principal,
interest, tax, and insurance,

M = mortgage loan amount,

An = amortization formula
n
th

= i / [1 1/(1+i) ], i =1/12 the annual
interest rate
and n = number of periodic payments,

Nevertheless, HMDA data do provide for the construction of a little-used but potentially powerful variable
that would enable analysts to more closely control for
the income leverage employed by the borrower and for
the associated payment risk when securing the reported
loan amount. This measure, when coupled with other
data reported under HMDA, could contribute greatly
to our understanding of the evolution, nature, and
Summary

(1)

so in the very-low- interest rate range of 3-, 2-, and

1-percent interest.

LTV = loan-to-home value ratio, M/V,

T
= annual property tax as percentage of
home value,

I
= annual home insurance premium as
percentage of home value.

With the aid of a couple key assumptions, one may use

the M/Y ratio together with the annual percentage rate
(APR) spread reported in the HMDA data to classify
loans generally according to the likely source of
financing and type of mortgage product used with its
associated payment risk. Returning to equation 2, if one
assumes that taxes and insurance amount to a specific
percentage (such as 1.5 percent) of home value and
lenders typically will not allow borrowers to exceed a
specific FER (such as 0.30), unless the FER is offset
with unusual compensations such as a very high credit
score,2 then the M/Y ratio is uniquely determined by
the qualifying interest rate. Thus, one may classify
loans in the HMDA file according to their status with
respect to high or low cost and high or low leverage.
High cost normally distinguishes nonprime loans from
prime loans and high leverage distinguishes loans with
temporary below-market qualifying advantages that
pose a risk of payment shock when rates adjust to the
fully indexed market rates or payments rise to retire
balances from initial negative amortization. Exhibit 2
shows that our demarcation between high- and low-cost
loans is approximated at an APR of 7 percent3 and
that high-leverage loans are distinguished with a step
function showing the maximum income qualifying

One may set LTV = 1 for simplicity and rearrange

terms to yield the following relationship showing that
the M/Y ratio is equal to the FER divided by 12 times
the sum of the amortization factor and monthly percentage contribution to tax and insurance:
M/Y = FER/ [12*(An + (T+I)/12]

(2)

The M/Y ratio for an interest-only mortgage is

obtained by substituting the monthly interest rate, i,
for An in equation 2.
As noted previously, one may directly observe in equation 2 that the M/Y ratio rises with a relaxation of the
FER or a reduction in the amortization factor An resulting from a reduction in the interest rate or increase in
the repayment period n. Exhibit 1 presents the M/Y
ratios for alternative qualifying interest rates and FERs
and confirms that the M/Y ratio rises with increases in
the FER at any given interest rate or with reductions
in the qualifying interest rate at any given FER. The
M/Y ratio is higher at every interest rate and FER combination for interest-only mortgages and dramatically

Exhibit 1. Mortgage-to-Income Ratio by Qualifying Interest Rate and Front-End Ratio

(30-Year Self-Amortizing Mortgage)
Qualifying
Interest Rate
20.0%
19.0%
18.0%
17.0%
16.0%
15.0%
14.0%
13.0%
12.0%
11.0%
10.0%
9.0%
8.0%
7.0%
6.5%
6.0%
5.5%
5.0%
4.5%
4.0%
3.5%
3.0%
2.0%
1.0%

Front-End Ratio
0.28

0.30

0.32

0.34

0.40

1.30
1.36
1.43
1.50
1.59
1.68
1.78
1.90
2.02
2.17
2.33
2.51
2.72
2.95
3.08
3.22
3.37
3.53
3.69
3.87
4.06
4.27
4.72
5.22

1.39
1.46
1.53
1.61
1.70
1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.30
3.45
3.61
3.78
3.96
4.15
4.36
4.57
5.05
5.60

1.48
1.56
1.63
1.72
1.81
1.92
2.04
2.17
2.31
2.48
2.66
2.87
3.11
3.37
3.52
3.68
3.85
4.03
4.22
4.43
4.65
4.88
5.39
5.97

1.58
1.65
1.74
1.83
1.93
2.04
2.16
2.30
2.46
2.63
2.83
3.05
3.30
3.59
3.74
3.91
4.09
4.28
4.49
4.70
4.94
5.18
5.73
6.34

1.86
1.94
2.04
2.15
2.27
2.40
2.54
2.71
2.89
3.09
3.32
3.59
3.88
4.22
4.40
4.60
4.81
5.04
5.28
5.53
5.81
6.10
6.74
7.46

Summary

Exhibit 2. Mortgage-to-Income Ratio at Varying Qualifying Interest Rates for Fully Indexed APR Rate
(30-Year Self-Amortizing Mortgage)
Qualifying
Interest
Rate
20.0%
19.0%
18.0%
17.0%
16.0%
15.0%
14.0%
13.0%
12.0%
11.0%
10.0%
9.0%
8.0%
7.0%
6.0%
5.0%
4.0%
3.0%
2.0%
1.0%

Fully Indexed Rate (APR)

20.0 19.0 18.0 17.0 16.0 15.0 14.0 13.0 12.0 11.0 10.0 9.0
%
%
%
%
%
%
%
%
%
%
%
%

8.0
%

7.0
%

6.0
%

5.0
%

4.0
%

1.39
1.46
1.53
1.61
1.70
1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

3.16
3.45
3.78
4.15
4.57
5.05
5.60

3.45
3.78
4.15
4.57
5.05
5.60

3.78
4.15
4.57
5.05
5.60

4.15
4.57 4.57
5.05 5.05 5.05
5.60 5.60 5.60 5.60

1.46
1.53
1.61
1.70
1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

1.53
1.61
1.70
1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

1.61
1.70
1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

1.70
1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

1.80
1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

1.91
2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

2.03
2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

2.17
2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

2.32
2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

2.49
2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

2.69
2.91
3.16
3.45
3.78
4.15
4.57
5.05
5.60

3.0
%

2.0
%

1.0
%

APR = annual percentage rate.

leverage that may be achieved for each percentage

point APR interval as one moves left from the lowest
to highest interval at 20 to 21 percent. Thus, any loan
with an APR falling between 6 and 7 percent would be
considered a high-leverage loan were it to have an M/Y
ratio greater than 3.45. Loans with an APR falling
between 7 and 8 percent and having an M/Y ratio greater
than 3.16 would be considered high-leverage loans,
and so on. The threshold ratios are somewhat conservative in that they are consistent with ratios that could
be achieved using an APR that is 2 percentage points
below the highest rate in each 1-point APR interval
using a more standard FER of 0.28 rather than 0.30.
(See Exhibit 1 and compare the M/Y ratio of 1.53 at
18-percent APR and 0.30 FER with the ratios at 19 and
17 percent APRs with an FER of 0.28.)

total loans, and the nearly three-fold increase in highleverage lending within the high-cost sector, from 3.7
to 9.7 percent, despite higher prevailing interest rates
in 2006. Avery, Brevoort, and Canner (2006), however,
advise caution in interpreting year-to-year changes in
the incidence of HMDA-reported high-cost lending
because flattening of the Treasury yield curve as occurred
through 2004 can result in a differentially increasing
proportion of APRs for adjustable-rate mortgage (ARM)
loans priced off shorter term rates to rise above the
high-cost reporting threshold rates in comparison to
fixed-rate loans.5 Despite the caution, Avery, Brevoort,
and Canner (2006) cite corroborating statistics from
Inside Mortgage Finance, indicating that, from 2004 to
2005, the share of higher cost, nonprime lending did
increase on the order of 7.5 percentage points, primarily
in the near-prime, Alt-A portion of the market.6

Using the M/Y ratio step factors in Exhibit 2 and the

positive APR spread indicator in the HMDA data,
loans originated4 in 2004, 2005, and 2006 were grouped
according to their status as high- or low-cost loans and
high- or low-leverage loans. The number and percentage of loans within each group are reported in the first
two panels of Exhibit 3.

More importantly, however, Exhibit 3 shows an increase

in the high-leverage lending share from 15.4 percent of
all loans in 2004 to 21.6 percent in 2006, accounting
for approximately 7.6 million loans over the 3-year
period regardless of high- or low-cost status. Furthermore, the third panel of Exhibit 3 shows that, although
refinancing declined somewhat over the period as a
proportion of total lending, it remained disproportionately higher as a percentage of high-leverage lending in
both the high- and low-cost sectors. Borrowers were

Perhaps the most striking observation in Exhibit 3 is

the more than doubling of the high-cost lending volume
from 2004 through 2006, from 14.5 to 29.4 percent of
Summary

Exhibit 3. Loan Distribution by High-Cost and High-Leverage Status

High Cost
Year

High Leverage

2004
2005
2006

505,109
1,243,047
1,218,190

2004
2005
2006

3.7
8.7
9.7

2004
2005
2006

68
60
60

Low Leverage

Low Cost
High Leverage

Number of Loans
1,491,222
1,613,215
2,522,268
1,534,849
2,468,539
1,497,403
Percent of Loans
10.8
11.7
17.7
10.8
19.7
11.9
Percent Refinancing
53
56
41
58
42
54

disproportionately much more likely to use high-leverage

loans in both the high- and low-cost sectors to refinance
rather than purchase their homes, which, on one level,
seems counterintuitive, because borrowers are in general most strapped and in need of maximum leverage
for their initial home purchase.

10,163,018
8,949,142
7,373,938

13,772,564
14,249,306
12,558,070

73.8
62.8
58.7

100.0
100.0
100.0

54
49
46

55
50
48

chases, they were significantly more likely to use highleverage loans within both the high- and low-cost markets
for home purchases. Moreover, Hispanic borrowers were
significantly more likely to engage in high-leverage
refinances in both the high- and low-cost markets.
Households in which the first listed name on the borrowers application was male (shown in panel 5) were
less likely to use high-leverage lending, particularly
high-cost leverage, to purchase or refinance their homes.
In addition, although owner-occupant borrowers (shown
in panel 6) accounted for roughly 85 percent of purchase
loans and 91 percent of refinance loans over the 3-year
period, they accounted for 95 to 97 percent of highleverage purchase and refinance loans in both the
high- and low-cost lending markets.

The disproportionate and increasing use of high-leverage

loans by refinancers is even more apparent in Exhibit 4.
High-leverage refinances increased from 16.6 percent
of 2004 refinance loans to 25.9 percent of 2006 refinance
loans. In comparison, high-leverage home purchase loans
increased from 14 percent of 2004 purchase lending to
17.8 percent of 2006 purchase lending. Moreover,
Exhibit 4 reveals a definite shift to the high-cost sector
in both purchase and refinance lending that continued
beyond 2005 well after mitigation of any reporting or
yield curve measurement problems that may have been
present in the 2004 data. High-cost loans accounted
for 29.0 percent of all purchase loans in 2006 compared
with 13.7 percent in 2004, and high-cost loans accounted
for 29.8 percent of all 2006 refinance loans compared
with 15.2 percent in 2004.

Panel 7 of Exhibit 4 shows that, although lower income

borrowers who have incomes of less than 80 percent of
area median income (AMI) accounted for roughly onequarter of both purchase and refinance lending, they
accounted for one-half or more of high- and low-cost
high-leverage lending in 2004. Despite these trends,
their share of overall lending declined by roughly 5
percent over the period from 2004 to 2006, and their
share of high-cost loans declined precipitously, particularly in the high-cost, high-leverage sector for both
purchase and refinance lending. The reduction in shares
for lower income borrowers was more than fully replaced
with the rising shares for higher income borrowers
(who have incomes greater than 120 percent of AMI,
shown in panel 9) in all lending sectors, particularly
high-cost and high-leverage purchase and refinance
lending. The $2,500 change in median income between
2004 and 2006 for high-cost, high-leverage refinancers
compared with the $1,900 change for refinancers overall (shown in panel 10) no doubt reflects the observed
shift from lower to higher income borrowers.

Exhibit 4 also presents demographic characteristics of

the borrowers using the various lending sectors to purchase or refinance their homes. The third panel of
Exhibit 4 shows that, although African-American borrowers accounted for approximately 7 to 9 percent of
purchase and refinance lending over the 3-year period,
they accounted for more than twice that percentage
(between 14 and 19 percent) of the high-cost market
for both purchase and refinance lending, with a slight
preference for greater use of high leverage to purchase
rather than refinance their homes.
Although Hispanic borrowers (shown in panel 4) were
also more likely to use high-cost loans for home pur

Total

Low Leverage

Summary

Exhibit 4. Characteristics of Purchase and Refinance Loans by High-Cost and High-Leverage Status
Purchase Loans
High Cost

Panel

Year

2004
2005
2006

2004
2005
2006

2.6
6.9
7.4

2004
2005
2006

19
16
19

2004
2005
2006

23
30
30

2004
2005
2006

57
58
57

2004
2005
2006

95
96
96

2004
2005
2006

51
36
32

2004
2005
2006

27
27
27

2004
2005
2006

22
37
41

10

2004
2005
2006

61,477
62,815
63,589

11

2004
2005
2006

55
57
57

Refinance Loans

Low Cost

High
Low
High
Low
Leverage Leverage Leverage Leverage

High Cost
Total

Low
High
Low
Leverage Leverage Leverage

Total

Number of Loans
4,709,799 6,281,960 343,024
796,465 897,896 5,453,219 7,490,604
4,605,148 7,235,261 745,142 1,031,307 893,602 4,343,994 7,014,045
3,988,388 6,576,043 732,972 1,048,668 814,837 3,385,550 5,982,027
Percent of Loans
11.1
11.4
75.0
100.0
4.6
10.6
12.0
72.8
100.0
20.6
8.9
63.6
100.0
10.6
14.7
12.7
61.9
100.0
21.6
10.4
60.7
100.0
12.3
17.5
13.6
56.6
100.0
African-American Borrower (%)
16
6
6
7
15
15
7
6
7
16
5
5
8
15
15
6
6
8
16
7
6
9
17
14
8
6
9
Hispanic Borrower (%)
19
15
10
12
15
9
15
8
9
21
14
9
13
18
10
16
8
10
23
13
10
15
18
12
16
9
12
Male Borrower (%)
63
63
69
67
54
62
60
69
66
63
63
68
66
55
62
61
68
65
62
62
68
65
54
61
59
66
63
Owner-Occupant Borrower (%)
84
96
84
86
97
91
97
91
92
81
95
81
83
97
89
97
90
91
78
96
83
84
96
86
97
89
90
Lower Income Borrower (Less Than 80 Percent of AMI) (%)
27
48
21
26
57
30
49
21
27
22
42
18
22
44
25
41
18
25
17
43
16
20
38
20
37
17
23
Middle-Income Borrower (Between 80 and 120 Percent of AMI) (%)
31
27
26
27
27
33
29
27
28
28
27
24
25
31
31
30
26
28
25
27
23
24
30
28
29
25
27
Higher Income Borrower (Greater Than 120 Percent of AMI) (%)
42
25
53
48
16
37
22
52
45
50
31
58
53
25
44
29
56
48
58
30
61
56
32
52
34
58
50
Borrower Median Income ($)
59,996 64,300
60,451
60,866 61,883
58,320 65,088
61,799
61,827
60,987 64,497
61,002
61,433 63,006
59,869 65,401
62,367
62,454
62,183 65,160
62,454
62,760 64,349
61,642 66,258
63,659
63,744
Underserved Area (%)
56
38
35
38
55
54
43
35
39
54
36
34
40
55
52
43
35
41
54
37
34
40
55
51
44
36
42

162,085
694,757 715,319
497,905 1,490,961 641,247
485,218 1,419,871 682,566

AMI = area median income.

Summary

High
Leverage

Low Cost

The first thing to note in Exhibit 5 is that, because the

mix of low- and high-cost borrowers within the various
(nongovernment-insured) lending sectors may change
from year to year and not all borrowers will use maximum leverage to purchase or refinance their homes,
year-to-year changes in the pattern of maximum leverage
provided may be best observed in the upper portions of
the distributions. Thus, one may compare the shaded
ratios above the 60th percentile for 2001, when prime
mortgage interest rates were in the neighborhood of
7 percent, with M/Y ratios in subsequent years to
observe how maximum leverage for purchase and refinance lending changed with interest rates and underwriting of various funding sources. Focusing on purchase
lending, one may observe that M/Y ratios increased by
30 to 40 basis points for loans funded by GNMA (or
Ginnie Mae), 35 to 55 basis points for loans funded by
government-sponsored enterprises (GSEs) Fannie Mae
and Freddie Mac, 35 to 70 basis points for loans funded
by portfolio lenders, and 35 to 60 basis points for loans
funded by private mortgage pool issuers. Returning to
Exhibit 1, one can observe that a 30-basis-point
increase from an M/Y ratio calculated at 7-percent
interest may be obtained by recalculating the ratio

Panel 11 of Exhibit 4 shows that, although overall

about 40 percent of purchase and refinance loans from
all lending sectors were made to borrowers living in
underserved areas, approximately 55 percent of high-cost,
high- and low-leverage lending was made to borrowers
purchasing and refinancing their homes in underserved
areas. This observation may not be particularly surprising
given the aforementioned relative shares of high-cost
lending for which minority and lower income borrowers
accounted. The relative constancy of that share over
the 3-year period despite the shift from lower income
to higher income borrowers discussed previously, however, is somewhat surprising.
Exhibits 3 and 4 clearly show that a significant shift to
higher leverage mortgage loans occurred between 2004
and 2006, particularly for borrowers purchasing or refinancing their homes with high-cost loans. What is not
yet clear is which funding sources were underwriting
the higher leverage loans and by how much leverage
was increasing for individual borrowers. Exhibit 5 presents the decile distribution of M/Y leverage ratios by
funding source for both purchase and refinance loans
originated in 2001 through 2006 and the average annual
interest rate prevailing in each year.

Exhibit 5. Decile Distribution of Mortgage-to-Income Leverage Ratios for 2001 Through 2006
Purchase
Funding
Year
Source
2001
2002
2003
GNMA
2004
2005
2006
2001
2002
2003
GSE
2004
2005
2006
2001
2002
Portfolio 2003
Lender 2004
2005
2006
2001
Private 2002
Mortgage 2003
2004
Pool
Issuer 2005
2006

Average
Annual
10th
20th
30th
40th
50th
60th
70th
80th
90th
Interest Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile
Rate
6.97%
1.46
1.74
1.96
2.16
2.35
2.56
2.78
3.05
3.46
6.54%
1.50
1.80
2.03
2.24
2.45
2.66
2.90
3.19
3.63
5.83%
1.56
1.88
2.13
2.36
2.58
2.81
3.07
3.38
3.86
5.84%
1.58
1.91
2.17
2.39
2.62
2.85
3.10
3.42
3.88
5.87%
1.58
1.92
2.18
2.40
2.63
2.86
3.12
3.43
3.88
6.41%
1.56
1.90
2.17
2.40
2.62
2.86
3.12
3.43
3.89
6.97%
1.13
1.45
1.68
1.89
2.10
2.31
2.56
2.87
3.35
6.54%
1.20
1.53
1.77
2.00
2.21
2.45
2.71
3.04
3.56
5.83%
1.26
1.60
1.87
2.11
2.35
2.61
2.90
3.27
3.83
5.84%
1.26
1.62
1.89
2.14
2.38
2.65
2.94
3.33
3.90
5.87%
1.28
1.64
1.91
2.17
2.42
2.69
3.00
3.39
3.97
6.41%
1.27
1.63
1.90
2.15
2.40
2.67
2.96
3.34
3.90
6.97%
0.34
0.64
1.08
1.44
1.75
2.04
2.35
2.70
3.22
6.54%
0.37
0.68
1.17
1.57
1.90
2.21
2.53
2.92
3.48
5.83%
0.41
0.78
1.31
1.72
2.07
2.40
2.75
3.16
3.76
5.84%
0.46
0.80
1.40
1.85
2.23
2.58
2.96
3.40
4.03
5.87%
0.40
0.61
0.88
1.50
2.00
2.42
2.82
3.27
3.89
6.41%
0.37
0.55
0.79
1.39
1.92
2.35
2.76
3.23
3.92
6.97%
0.47
0.96
1.36
1.67
1.92
2.18
2.46
2.79
3.28
6.54%
0.45
0.91
1.43
1.77
2.06
2.34
2.64
3.00
3.52
5.83%
0.53
1.03
1.58
1.94
2.25
2.55
2.88
3.26
3.84
5.84%
0.53
0.83
1.48
1.92
2.27
2.60
2.95
3.35
3.93
5.87%
0.51
0.74
1.19
1.80
2.20
2.56
2.92
3.33
3.88
6.41%
0.46
0.66
0.92
1.62
2.06
2.41
2.76
3.14
3.65

Summary

Exhibit 5. Decile Distribution of Mortgage-to-Income Leverage Ratios for 2001 Through 2006 (continued)
Refinance
Average
Funding
Annual
10th
20th
30th
40th
50th
60th
70th
80th
90th
Year
Source
Interest Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile
Rate
2001 6.97%
1.21
1.50
1.73
1.94
2.15
2.38
2.63
2.97
3.50
2002 6.54%
1.09
1.40
1.64
1.86
2.08
2.32
2.60
2.95
3.52
2003 5.83%
1.01
1.31
1.54
1.76
2.00
2.24
2.53
2.91
3.50
GNMA
2004 5.84%
1.05
1.40
1.67
1.91
2.16
2.43
2.75
3.14
3.75
2005 5.87%
1.08
1.44
1.74
2.00
2.26
2.53
2.84
3.21
3.77
2006 6.41%
1.45
1.77
2.03
2.27
2.50
2.75
3.02
3.35
3.82
2001 6.97%
1.04
1.31
1.53
1.73
1.94
2.17
2.44
2.78
3.31
2002 6.54%
1.01
1.28
1.51
1.72
1.94
2.18
2.46
2.82
3.38
2003 5.83%
0.97
1.25
1.48
1.70
1.92
2.17
2.47
2.86
3.47
GSE
2004 5.84%
1.07
1.38
1.64
1.88
2.14
2.42
2.75
3.18
3.85
2005 5.87%
1.21
1.57
1.86
2.14
2.43
2.74
3.09
3.55
4.23
2006 6.41%
1.24
1.61
1.92
2.22
2.51
2.83
3.18
3.63
4.29
2001 6.97%
0.40
0.69
0.99
1.27
1.54
1.83
2.15
2.55
3.14
2002 6.54%
0.48
0.81
1.10
1.38
1.65
1.93
2.26
2.68
3.31
Portfolio 2003 5.83%
0.53
0.86
1.14
1.40
1.68
1.98
2.33
2.78
3.46
0.52
0.88
1.24
1.59
1.95
2.33
2.76
3.28
4.02
Lender 2004 5.84%
2005 5.87%
0.43
0.71
1.07
1.48
1.90
2.33
2.80
3.34
4.08
2006 6.41%
0.38
0.60
0.86
1.22
1.67
2.15
2.65
3.21
3.94
2001 6.97%
0.75
1.18
1.48
1.74
2.00
2.25
2.55
2.91
3.46
Private 2002 6.54%
0.91
1.29
1.58
1.83
2.09
2.37
2.68
3.07
3.66
Mortgage 2003 5.83%
0.98
1.36
1.65
1.92
2.19
2.48
2.82
3.24
3.88
2004 5.84%
1.04
1.55
1.91
2.23
2.54
2.87
3.24
3.69
4.34
Pool
0.91
1.57
2.00
2.36
2.71
3.06
3.45
3.90
4.55
Issuer 2005 5.87%
2006 6.41%
0.62
1.18
1.75
2.18
2.56
2.92
3.31
3.75
4.38
GNMA = Ginnie Mae. GSE = government-sponsored enterprise.
Note: Average annual interest rate is from the Freddie Mac Series of Monthly Average Commitment Rate and Points on 30-Year
Fixed-Rate Mortgages.

using an interest rate of 6 percent, 100 basis points

below 7 percent. A 60-basis-point increase in the ratio
is equivalent to a 200-basis-point reduction to an
interest rate of 5 percent, and a 90-basis-point increase
in the M/Y ratio is obtained with a 300-basis-point
reduction in the interest rate. Thus, as average annual
interest rates declined by roughly 120 basis points over
the period, from 7 to 5.84 percent, maximum M/Y
leverage ratios for purchase loans rose to levels corresponding to a 200-basis-point reduction in interest in
all but the GNMA sector, indicating a widespread general easing of underwriting in the conventional markets
for high-leverage lending.

90-basis-point increase in the leverage ratio for a 6- to

7-percent interest rate loan is equivalent to a 300-basispoint reduction, to 3 or 4 percent, in the interest rate.
Again, because the decline in average annual interest
over the period was about 120 basis points and was
only 50 basis points below 2001 levels in 2006, it appears
that a substantial easing of underwriting occurred in
the conventional markets for high-leverage refinance
lending that went well beyond that observed for home
purchase lending.
Exhibit 6 shows the distribution of spreads between
home purchase M/Y ratios and refinance M/Y ratios by
decile from 2001 through 2006. Observing the unshaded
cells of the exhibit, one may note that the spread is
generally positive (or near zero where negative) for all
funding sources except private mortgage pool issuers
and the GSEs in 2005 and 2006, indicating that, in
general, higher income leverage was a necessity and
was granted more often under standard underwriting
practices when stretching for an initial home purchase.

Moreover, Exhibit 5 shows that maximum M/Y leverage

ratios for refinance loans increased by 16 to 38 basis
points for loans funded by GNMA, 60 to 95 basis points
for loans funded by the GSEs Fannie Mae and Freddie
Mac, 50 to 90 basis points for loans funded by portfolio
lenders, and 70 to 90 basis points for loans funded by
private mortgage pool issuers. As noted previously, a
Summary

10

Exhibit 6. Decile Distribution of Spread Between Purchase and Refinance Mortgage-to-Income Ratios for 2001
Through 2006
Purchase M/YRefinance M/Y
Average
Funding
Annual
10th
20th
30th
40th
50th
60th
70th
80th
90th
Year
Source
Interest Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile Percentile
Rate
2001
2002
2003
GNMA
2004
2005
2006
2001
2002
2003
GSE
2004
2005
2006
2001
2002
Portfolio 2003
Lender 2004
2005
2006
2001
Private 2002
Mortgage 2003
2004
Pool
Issuer 2005
2006

6.97%
6.54%
5.83%
5.84%
5.87%
6.41%
6.97%
6.54%
5.83%
5.84%
5.87%
6.41%
6.97%
6.54%
5.83%
5.84%
5.87%
6.41%
6.97%
6.54%
5.83%
5.84%
5.87%
6.41%

0.25
0.41
0.54
0.53
0.50
0.11
0.09
0.19
0.29
0.20
0.06
0.03
0.06
0.11
0.12
0.06
0.03
0.01
0.28
0.46
0.45
0.51
0.39
0.16

0.24
0.40
0.58
0.51
0.47
0.13
0.14
0.24
0.36
0.24
0.06
0.01
0.05
0.12
0.08
0.08
0.10
0.04
0.22
0.38
0.33
0.71
0.83
0.52

0.23
0.40
0.59
0.50
0.44
0.14
0.16
0.27
0.39
0.25
0.05
0.02
0.09
0.07
0.17
0.15
0.18
0.06
0.12
0.14
0.07
0.43
0.81
0.83

0.22
0.38
0.59
0.48
0.40
0.13
0.16
0.28
0.42
0.25
0.03
0.07
0.17
0.19
0.32
0.26
0.02
0.17
0.07
0.06
0.02
0.30
0.57
0.56

0.20
0.36
0.58
0.46
0.37
0.12
0.15
0.28
0.43
0.24
0.01
0.11
0.21
0.25
0.39
0.28
0.10
0.25
0.08
0.03
0.06
0.27
0.51
0.51

0.18
0.34
0.57
0.41
0.33
0.11
0.14
0.27
0.43
0.22
0.04
0.16
0.21
0.27
0.42
0.25
0.08
0.19
0.07
0.03
0.07
0.27
0.50
0.51

0.15
0.30
0.54
0.35
0.28
0.09
0.12
0.26
0.43
0.19
0.09
0.22
0.20
0.27
0.42
0.20
0.02
0.11
0.09
0.04
0.06
0.29
0.53
0.55

0.09
0.24
0.48
0.28
0.22
0.09
0.09
0.23
0.41
0.15
0.16
0.30
0.15
0.24
0.38
0.12
0.08
0.02
0.12
0.07
0.03
0.34
0.57
0.61

0.04
0.11
0.36
0.13
0.11
0.07
0.04
0.17
0.37
0.05
0.26
0.38
0.08
0.17
0.30
0.02
0.20
0.03
0.18
0.14
0.04
0.41
0.67
0.72

GNMA = Ginnie Mae. GSE = government-sponsored enterprise. M/Y = mortgage-to-income ratio.

els of leverage were permitted to stretch well beyond

limits applied to those purchasing homes.

It is interesting to note, however, that the spreads for

GSEs and private mortgage pool issuers switch dramatically in 2005 and 2006 from being positive for
GSEs and near zero for private pool issuers to negative
as the maximum leverage granted to refinancers
increased by 50 to 60 basis points compared with that
granted to purchasers.

Exhibit 7 shows the state-by-state variation and

growth over time of the high-leverage share of all
lending and the refinance portion of that high-leverage
lending across the United States from 2004 through
2006. In 2004, high-leverage lending accounted for less
than 11 percent of all lending in nearly one-half of the
states (shown with the lightest shading), while, by
2006, the most lightly shaded states where the share
of high-leverage lending was less than 11 percent had
declined to only eight states. The maps also show that
the highest income leverage was granted on the west
coast and in the northeastern states, with California
having the highest occurrence of high-leverage lending
of all the lower 48 states: 29, 37, and 36 percent in
2004, 2005, and 2006, respectively.7 High-leverage
lending was also prevalent in high-cost areas such as
Washington, Nevada, and Arizona in the West and
Massachusetts, Rhode Island, Maryland, the District

Thus, it would appear that much of todays current

mortgage turmoil was triggered by widespread easing
of the underwriting standards in conventional lending
sectors that permitted borrowers to leverage their
incomes beyond standards prevailing in 2001 and earlier, using qualifying advantages of short duration,
such as ARMs with low teasers or other variations of
ARMS with low introductory payments. Moreover, it
appears that the relaxation of standards was most
egregious in the refinancing sector, particularly among
private mortgage pool issuers, where borrowers who
already owned homes and whose circumstances would
have permitted refinancing at considerably lower lev

11

Summary

Exhibit 7. Mortgage Leverage and Refinance Activity in the United States

2004

2005

2006

20042006

57
0
.6

0.

47
0
.5

7
0.

rates expire. As the temporary qualifying interest rate

expires and reprices to market, the mortgage payment
and the associated new FER is carried well above the
original qualifying payment-to-income ratio in order to
remain consistent with the M/Y leverage embodied in
the mortgage contract. (Note in Exhibit 1 that for a
loan originally underwritten at 3.5 percent interest and
a FER of 0.30, the FER must rise to 0.40 to preserve
the same M/Y leverage of 4.36 as the interest rate rises
to market at 6.5 percent.) The map for 2006 and the
consolidated map reflecting lending for all 3 years are
in reasonably close alignment with maps and information based on other data that show where delinquency
and foreclosure problems have been greatest.8

Thus, the more heavily shaded areas of the maps identifying states where high-leverage lending activity has
been greatest also indicate where the potential for repayment problems and subsequent foreclosure problems
are likely to be highest as low introductory interest
Summary

37
0
.4

of Columbia, and Virginia in the East. High-leverage

refinance activity in the lower 48 states was greatest
in Rhode Island in 2004 and 2005 and greatest in
California in 2006. Other states in which refinance
activity constituted more than 57 percent of high-leverage
lending in 2006 include Maine, New Hampshire,
Vermont, Massachusetts, Rhode Island, New Jersey,
Maryland, the District of Columbia, Florida, Michigan,
Wisconsin, Illinois, Arizona, and Hawaii.

0.

0.

27
0
.3

9
32
0
.3

Refinance Share of High-Leverage Landing

0.

25
0
.3

0.

0.1
8
0.
2

0.

04

.11
0.1
1
0.1
8

High-Leverage Share of Lending

12

point reduction below market interest when setting a

demarcation threshold between high and low leverage.

The preceding analysis indicates that a substantial

weakening of underwriting standards, particularly for
refinancing, and a significant shift to higher leverage
mortgage lending occurred between 2004 and 2006,
particularly for borrowers purchasing or refinancing
their homes with high-cost loans. Moreover, borrowers
were disproportionately much more likely to use highleverage loans in both the high- and low-cost sectors to
refinance rather than purchase their homes. Although
minority and low-income borrowers were disproportionately more likely to use high-leverage lending in
the high-cost sector to purchase and refinance their
homes, higher income borrowers were also well represented in high-leverage lending in both high- and lowcost sectors. Certainly, more careful and controlled
analysis is necessary; however, it does not appear on
the face of it that evidence supports the assertions of
some that limited government policies advancing
homeownership or affordable housing goals are responsible for the current mortgage market turmoil. The
causes are likely more complex and broad in nature.

Loans are actually classified high cost if a positive APR

spread is reported in HMDA data. Analysis of HMDA
APR spreads of 3 percentage points or more over prevailing Treasury rates and Freddie Mac commitment rates
pointed to an approximate dividing line of 7 percent.
3

The loans analyzed in this article are HMDA purchase

and refinance loans for one- to four-family homes (excluding manufactured homes).
4

See Avery, Brevoort, and Canner (2006: A141-A152).

Ibid, A-144.

High-leverage lending (and high-leverage refinancing)

activity in Hawaii and Alaska (not shown in Exhibit 7)
was as follows: Hawaii31, 37, and 39 percent (and 61,
62, and 60 percent); Alaska13, 16, and 22 percent (and
37, 38, and 41 percent) in 2004, 2005, and 2006, respectively.
7

See http://www.newyorkfed.org/newsevents/news/
regional_outreach/2008/an080401.html and http://www.
newyorkfed.org/mortgagemaps/current/.
8

Notes
The authors, William J. Reeder and John P. Comeau,
thank Ismail Mohamed, Jian Zhou, and John Mubiru for
valuable assistance with this project.
1

Reference
Avery, Robert B., Kenneth P. Brevoort, and Glenn B.
Canner. 2006. Higher-Priced Home Lending and the
2005 HMDA Data, Federal Reserve Bulletin 92
(September): A123A166.

With automated underwriting, lenders have been more

willing to relax the FER requirement, which is why we
subsequently choose 0.30 rather than 0.28, which was traditional for years, and allow for as much as a percentage
2

13

Summary